Treatment in an ICU staffed by appropriately trained intensivist clinicians improves survival in critically ill patients. Yet many patients lack access o this level of critical care, particularly in small hospital and rural geographic areas, leading to excess mortality and creating significant socioeconomic disparities. To address this problem many hospitals have adopted ICU telemedicine, a health care delivery innovation which uses audio---visual technology to provide critical care services from a distance. By expanding access to high-quality critical care, ICU telemedicine has great potential to improve survival in critical illness. However, the success of telemedicine is hindered by critical deficiencies in our understanding of how and where this technology is best applied. Telemedicine is a complex intervention, and existing programs vary significantly in both the components of intervention and the clinical setting in which it is used. In turn, studies of telemedicine demonstrate mixed result, with some showing a large mortality benefit and others showing no benefit or even suggesting harm. As a result, decision makers have little guidance regarding how and where to use this potentially transformative technology, if at all. In this project we will address this knowledge ga by identifying the key clinical and organizational factors associated with ICU telemedicine effectiveness. Our central hypothesis is that objective, measurable, clinical and organizational factors will distinguish effective programs from ineffective ones. First, we will use national data on Medicare beneficiaries and an innovative risk--- adjustment procedure to quantify hospital---level variation in the impact of ICU telemedicine on patient mortality, ranking each hospital that has adopted ICU telemedicine from most effective to least effective. Second, we will conduct site visits at five hospitals with the greatest telemedicine effect and five hospitals with least telemedicine effect; along with two site visits at hospitals that have stopped using ICU telemedicine; performing in-depth qualitative analyses to identify the clinical and organizational factors associated with ICU telemedicine effectiveness. Third, we will develop and field a survey of all US hospitals that have adopted ICU telemedicine, linking the survey back to patient---level outcomes data and quantitatively defining ICU telemedicine clinical and organizational best practices. By shifting the paradigm of ICU telemedicine evaluation away from whether it works to how and where it works best, this project will be the first rigorous examination of the factors that define successful ICU telemedicine implementation. Ultimately, these results will provide clinicians, hospital administrators with immediate, actionable data about how to use ICU telemedicine most effectively and efficiently, directly leading to improved access to critical care and improved survival for critically ill patients.